Bug #14131: test_backtrace.rb: Fails on ppc64le

• Status: Closed
• Assignee: -
• Target version: -
• Backport: 2.3: UNKNOWN, 2.4: UNKNOWN

[leitao (Breno Leitao)]

Current test_backtrace.rb fails on ppc64el environment with the following stack:

unstable ➜  ruby git:(trunk) ✗ ./miniruby  -I./lib -I. -I.ext/common  ./tool/runruby.rb --extout=.ext  -- --disable-gems "./test/runner.rb" --ruby="./miniruby -I./lib -I. -I.ext/common  ./tool/runruby.rb --extout=.ext  -- --disable-gems" --excludes-dir=./test/excludes --name=test_caller_lev  test/ruby/test_backtrace.rb
Run options: "--ruby=./miniruby -I./lib -I. -I.ext/common  ./tool/runruby.rb --extout=.ext  -- --disable-gems" --excludes-dir=./test/excludes --name=test_caller_lev

# Running tests:

[1/1] TestBacktrace#test_caller_lev = 0.00 s
  1) Error:
TestBacktrace#test_caller_lev:
SystemStackError: stack level too deep
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:93:in `block (2 levels) in test_caller_lev'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:92:in `times'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:92:in `block in test_caller_lev'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:93:in `block (2 levels) in test_caller_lev'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:92:in `times'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:92:in `block in test_caller_lev'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:93:in `block (2 levels) in test_caller_lev'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:92:in `times'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:92:in `block in test_caller_lev'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:93:in `block (2 levels) in test_caller_lev'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:92:in `times'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:92:in `block in test_caller_lev'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:93:in `block (2 levels) in test_caller_lev'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:92:in `times'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:92:in `block in test_caller_lev'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:93:in `block (2 levels) in test_caller_lev'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:92:in `times'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:92:in `block in test_caller_lev'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:93:in `block (2 levels) in test_caller_lev'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:92:in `times'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:92:in `block in test_caller_lev'
    /home/breno/ruby/ruby/test/ruby/test_backtrace.rb:106:in `block in test_caller_lev'

Finished tests in 0.005963s, 167.7020 tests/s, 1341.6160 assertions/s.
1 tests, 8 assertions, 0 failures, 1 errors, 0 skips

ruby -v: ruby 2.5.0dev (2017-11-27 trunk 60922) [powerpc64le-linux]

Also, the bisect points to the following commit:

 commit c4e2cf466448f4283fd3f8a17a73f5fa9b745fe1
 Author: normal <normal@b2dd03c8-39d4-4d8f-98ff-823fe69b080e>
 Date:   Thu Jun 8 20:58:03 2017 +0000

     tool/runruby.rb: test with smallest possible machine stack
    
     Lets ensure none of our C functions use too much stack space and
     fix all excessive stack usage before releasing the next version.
     Reducing C stack usage should reduce conservative GC scanning
     time and improve performance.
    
     If there are platform-dependent test failures; excessive stack
     usage should be fixed; rather than increasing minimum values or
     removing these envs from testing.

Looking at the code, I also found that align the stack to 64kb (size of the page in ppc64le) solves the problem also, something like:

Author: Breno Leitao <[e-mail]>
Date:   Mon Nov 27 10:47:00 2017 -0500

    stack: align to 64kb
    
    This is a workaround for ppc64el which uses 64kb page size. This align
    the stack to the page size. It shouldn't have issues with amd64, but I
    am not the expert.
    
    Signed-off-by: Breno Leitao <[e-mail]>

diff --git a/vm_core.h b/vm_core.h
index 41663fd43e..71f5ee5ba2 100644
--- a/vm_core.h
+++ b/vm_core.h
@@ -576,7 +576,7 @@ typedef struct rb_vm_struct {
 
 /* default values */
 
-#define RUBY_VM_SIZE_ALIGN 4096
+#define RUBY_VM_SIZE_ALIGN 4096 * 16
 
 #define RUBY_VM_THREAD_VM_STACK_SIZE          ( 128 * 1024 * sizeof(VALUE)) /*  512 KB or 1024 KB */
 #define RUBY_VM_THREAD_VM_STACK_SIZE_MIN      (   2 * 1024 * sizeof(VALUE)) /*    8 KB or   16 KB */

Anyway, any idea on how to solve this issue?

This bug showed up initially in the Debian as https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=881772

[vo.x (Vit Ondruch)]

• Is duplicate of Bug #13757: TestBacktrace#test_caller_lev segaults on PPC added

[vo.x (Vit Ondruch)]

• Status changed from Open to Closed

Hi Breno, this is duplicate of #13757. Could you please check the discussion there? I didn't have enough free cycles lately to continue to dig into this ..

[leitao (Breno Leitao)]

I was able to narrow down this issue to the following code:

 max = 20
 rec = lambda{|n|
	
   if n > 0
       rec[n-1]
   end
 }

 #rec[max]
 Fiber.new{
   rec[max]
 }.resume

That you should call with:

 # ./miniruby tool/runruby.rb  lambda.rb

Some further findings:

• The problem is not reproducible if we use functions recursion instead of lambda recursion.

• The problem does not happen if we run the recursion outside of a Fiber.